South Africa has pushed back the launch of its national artificial intelligence (AI) policy to January 2027. This decision was made after the government had to pull its first draft because it contained fake academic references a problem caused by the very technology it was trying to regulate.
South Africa has pushed back the launch of its national artificial intelligence (AI) policy to January 2027. This decision was made after the government had to pull its first draft because it contained fake academic references a problem caused by the very technology it was trying to regulate.
In April 2026, the government released a draft document designed to guide how South Africa would use and control AI safely. However, journalists and researchers soon discovered a major problem: the list of sources at the end of the document included fake academic studies and imaginary journals.
Experts call this an “AI hallucination.” It happens when an AI tool is used to write text and makes up facts or sources that sound completely real but do not exist.
The Government’s Response
Communications Minister Solly Malatsi called the mistake a “massive oversight.” He admitted that the department failed to check the sources properly before making the document public.
To fix the issue, the government has taken several steps:
• Suspensions: Two officials involved in writing the policy have been suspended.
• Expert Team: The government hired a new panel of independent tech, law, and data experts to rewrite the policy from scratch.
• Better Checks: The department promised to create stricter internal rules so AI cannot be used blindly in government work again.
Why This Delay Matters
South Africa wanted to be a leader in Africa for digital technology and AI rules. This scandal is an embarrassing setback.
While the government restarts its work, local businesses and tech developers will have to wait another nine months to get clear guidance on what is allowed. The new timeline means the fixed rules will not be ready for the public to read until January 2027.
The Irony of the Scandal
This situation highlights a lesson that tech experts have been sharing for years: never trust AI blindly.
Ironically, South Africa’s draft policy was supposed to teach citizens about the dangers of using AI without human supervision. By failing to check its own work, the government accidentally provided the perfect example of why those rules are needed.
South Africa Pulls the Plug on Draft AI Policy Over Fake AI Citations
This news broadcast provides an overview of how fake, AI-generated citations embarrassed the South African government and forced them to withdraw their initial technology policy.
